Uncertainty quantification in fine-tuned LLMs using LoRA ensembles
Balabanov, Oleksandr, Linander, Hampus
Fine-tuning large language models can improve task specific performance, although a general understanding of what the fine-tuned model has learned, forgotten and how to trust its predictions is still missing. We derive principled uncertainty quantification for fine-tuned LLMs with posterior approximations using computationally efficient low-rank adaptation ensembles. We analyze three common multiple-choice datasets using low-rank adaptation ensembles based on Mistral-7b, and draw quantitative and qualitative conclusions on their perceived complexity and model efficacy on the different target domains during and after fine-tuning. In particular, backed by the numerical experiments, we hypothesise about signals from entropic uncertainty measures for data domains that are inherently difficult for a given architecture to learn.
Feb-19-2024
- Country:
- Europe (0.68)
- North America > United States
- Minnesota > Hennepin County > Minneapolis (0.14)
- Genre:
- Research Report (0.64)
- Industry:
- Education (0.48)